Interpretation of the total power absorbed calculated with the solar generation analysis group


Hi everybody,

I am running simulations with the FDTD software, and I need to analyse the plot of the total absorbed power vs wavelength (ploted from the solar generation analysis group script). I don’t know how to interpret it clearly:

  • What is the unit of the total power absorbed ploted? (It is written in the script “fraction of source power”, I don’t really know what it means)
  • Is there a kind of normalization that I have to take into account in the interpretation of the value of the power? (I am using the CW normalization for the source) Is it wrong if the value of the power goes beyond 1? Is it wrong if it is negative?
  • What is the influence of the volume of the FDTD region on the value of the power?
  • What is the influence of the downsampling on the accuracy of the power calculation?

Thank you in advance for your reply!


The quantity “total absorbed power” reported by the analysis group is unitless since it is the ratio of absorbed power to the total input power (Watt / Watt). The result basically tells you what fraction of the input optical power gets absorbed by the device at each wavelength.

Given the description of “total absorbed power”, you can see that the normalization was done w.r.t. the source power. Since the value gives you the “fraction of source power” absorbed in the device, any negative or greater than 1 value is not correct. If you get such values then the simulation setup should be checked. That said, in devices where you have resonant behavior (such as a plasmonic structure to enhance absorption), the value at the resonant wavelength can often be greater than 1 and it can be quite difficult to find the correct simulation setings to resolve this issue.

As long as you are simulating over a unit cell, the volume of the simulation region should not have any effect on this value.

The solar generation analysis group calculates the absorbed power by using the electric field data recorded by a DFT monitor and the index data recorded by an index monitor. If downsampling is used, the monitors records the data at fewer grid points (reduced by a factor equal to the downsampling parameter). Thus using downsampling can reduce the accuracy of the result if the monitors record the data on a grid that is too coarse.


Now I understand better, thank you very much for yuour reply!